The board deck shows plant OEE at 84% and trending up. On the roll-up, Line 2 — the constraint — sits at 66%, and its lowest factor is Performance. Corporate asks where the plant's focus is this quarter.
Q1 What's the answer that reads the page correctly?
A
Speed and minor-stop losses on Line 2 — the constraint's weakest factor names the loss family that gates the schedule. Drill the worst contributor, then its lowest factor. Performance low means slow cycles and minor stops, and on the constraint that loss is plant throughput.
B
Plant-wide quality initiatives — defects hurt every line at once. Nothing on the page points to Quality. Focus chosen by instinct instead of the roll-up is how attention lands where it changes the least money.
C
Nowhere specific — at 84% and rising, the plant should hold course. The headline hides a constraint running 18 points below it. A high plant OEE can still hide one line bleeding the schedule — that's why you read top-down.
Covered in Module 01
Quarter close: plant OEE improved three points, but the verified-savings ledger barely moved.
Q2 How do you read that page?
A
Normal lag — ledgers always trail, so report the OEE win and move on. 'Probably lag' is a hypothesis to check, not a reason to stop reading. A flat ledger under rising OEE is exactly the gap the one-page view exists to expose.
B
The gains aren't reaching the P&L — the OEE-to-ledger gap is itself the finding to investigate. Losses on the left, banked recoveries on the right. When the left improves and the right doesn't, something between action and verification is broken — that's your next question.
C
The OEE numbers must be inflated — recalculate them before reporting. Both numbers can be honest at once. The gap isn't an error to reconcile away; it's a signal about where improvement is stalling before it banks.
Covered in Module 01
Sales asks if the plant can absorb a new recurring order. Your line lead says the floor is 'maxed out.' The page says OEE 80%, TEEP 55% on the lines involved.
Q3 What does the page say back?
A
Not maxed out — a 25-point OEE-to-TEEP gap is idle calendar time. The order can fit in hours you already own. 'Maxed out' is a claim; TEEP is the arbiter. TEEP well below OEE means you have unscheduled time before you have a true constraint.
B
Accept it only if new equipment is approved first. Buying capacity while roughly a third of the calendar sits unscheduled is paying for a problem you don't have.
C
Decline the order — the lead knows the floor better than a metric does. Floor knowledge matters for how to schedule the hours, but whether spare hours exist is exactly what TEEP settles — and 55% says there's room.
Covered in Module 01
Ten minutes into the daily tier board, a supervisor pulls up a personal spreadsheet to argue that yesterday's downtime total is wrong.
Q4 What's the right move for the meeting?
A
Task an analyst with reconciling the platform against his spreadsheet each morning. That makes data archaeology a permanent job. The point of one platform is that there's nothing to reconcile.
B
Let both numbers stand and split the difference for the record. Averaging a logged number with a private one corrupts the record and licenses everyone to bring their own spreadsheet next time.
C
Point at the platform's reason-code log — the single source — and return to decisions and owners. If a code is wrong, fix the logging at the source afterward. Debating whose number is right is one of the two failure modes the fixed agenda exists to kill. Data disputes go to the logging discipline, not the meeting clock.
Covered in Module 02
The weekly review convenes an hour after a dramatic breakdown on Line 1. The room wants to spend the session on it. Last week's three actions sit unchecked on the board.
Q5 How does the review open?
A
With last week's three actions — the breakdown already belongs to the shift and daily tiers, and a review that skips the loop teaches owners that commitments evaporate. Open by closing. The live event is being handled at the altitude that owns it; the weekly's job is trends, backlog, and follow-through.
B
With the breakdown — it's the most urgent thing in the plant right now. Urgent and live means it's the lower tiers' work already underway. Re-litigating it at the weekly duplicates effort and starves the follow-through the weekly exists for.
C
With a poll for new issues, then the breakdown, then old actions if time allows. 'If time allows' is how actions rot. The opening slot is the enforcement mechanism — lose it and the cadence becomes theater.
Covered in Module 02
A vendor pitches a $200K/yr improvement for the constraint line. When you ask how the saving would be measured into the verified ledger, they offer a case study from another plant instead.
Q6 What's the disciplined response?
A
No measurement plan, no funding — if it can't verify in the ledger, its savings can't be treated as real. The ledger rule disciplines vendors and pet projects alike: define the before/after measurement first, or the $200K stays a brochure number.
B
Fund it — $200K dwarfs everything on the ranked list, and it's on the constraint. Constraint location is necessary, not sufficient. An unmeasurable claim can never be proven to have worked — you'd be buying a story.
C
Fund a half-scale pilot to reduce the risk. Half of unmeasurable is still unmeasurable. A pilot only helps if its verification method is defined before the spend.
Covered in Module 03
Two changeover fixes compete for the same crew: $6,000/wk recoverable on Line 1, which has schedule slack, and $4,000/wk on Line 3, which gates plant throughput.
Q7 Where does the crew go?
A
Alternate weeks between the two lines to keep both leads happy. Fairness to lines isn't the objective — banked savings and throughput are. Splitting effort on the constraint fix just delays its verification.
B
Line 3 — recovered hours on the constraint bank as throughput; the bigger number on a non-constraint line mostly buys idle time. An hour saved where the schedule was never waiting often banks nothing. You read the ranked dollars and the constraint position together.
C
Line 1 — $6,000 beats $4,000, and the ranking is there for a reason. The ranking surfaces recoverable value; you weight it toward the constraint. A slack line converts recovered time into waiting, not money.
Covered in Module 03
A capex request argues: 'Whenever this line is scheduled, it runs nearly flat out — OEE 82%. We need a second one.' The line's TEEP reads 44%.
Q8 What did the request miss?
A
That OEE grades only scheduled hours. At TEEP 44% vs OEE 82%, nearly half the calendar is unscheduled — the cheapest new capacity is hours already owned. High OEE with low TEEP is the signature of a well-run, under-scheduled asset. You don't buy a second machine to solve a scheduling decision.
B
That the line should push OEE to 90% before any request is considered. Chasing the last OEE points on an under-scheduled line is the wrong sequence — the idle calendar time dwarfs what's left inside OEE.
C
Nothing — 82% OEE proves the asset is exhausted, so the request stands. OEE says how well you run while scheduled; it says nothing about the hours you didn't schedule. That's exactly the blind spot TEEP exists to expose.
Covered in Module 04
Last week Line 6 failed to produce for 60 hours: 38 unscheduled, 14 down for breakdowns during scheduled time, 8 lost to running below rated speed.
Q9 Which lever addresses the biggest bucket?
A
A reliability and changeover program — breakdowns are the real enemy. That lever fits the 14 down hours — the second bucket. Aiming the biggest effort at the second-biggest loss is how idle time survives untouched.
B
A capital request — a faster machine claws back the slow-running hours. The 8 slow hours are a Performance loss — speed and minor-stop work, not new iron. Capex is the last resort, and this line isn't near it.
C
Scheduling — 38 of the 60 hours are calendar time nobody planned to use. Fill them or sell the capacity; no capex, no engineering. Each kind of 'not running' has its own lever. The unscheduled bucket is the largest and the cheapest: it's a planning decision, not a plant problem.
Covered in Module 04
Demand on your constraint product is up 15% and holding. You need capacity.
Q10 What's the sequence?
A
File the capex now — lead times are long — and run the improvement work in parallel. 'Order now, improve later' commits capital before the constraint is proven. If the recovered hours cover the 15%, you've bought an asset to solve a schedule.
B
Schedule the idle calendar time, then recover OEE losses on the constraint, then request capex only if TEEP nears its ceiling with demand still unmet. Recovered capacity before bought capacity, in that order. Equipment funded while recoverable hours sit on the floor runs at the same loss rate as the asset you already have.
C
Raise OEE across all lines first so the whole plant contributes. Non-constraint OEE gains mostly bank nothing — the schedule isn't waiting on those lines. The sequence targets the constraint, not the average.
Covered in Module 04
An engineer's fix visibly improved a line — everyone on the floor agrees it runs better. But no baseline was captured, so there's no before/after to measure. The quarterly savings number is being assembled.
Q11 What happens to this improvement?
A
Reconstruct a baseline from memory and log the delta against it. A remembered baseline is an estimate wearing a measurement's clothes. The ledger's value is that nothing in it is reconstructed.
B
It stays out of the ledger — uncounted — and the lesson is to capture the baseline before the next fix ships. 'If it can't be measured into the ledger, it isn't counted' has no goodwill exception. The rule stings exactly once per team — then baselines get captured.
C
Book a conservative estimate — the improvement is real and visible to everyone. Visible-to-the-floor is not traceable-to-finance. One estimated entry means every ledger number now needs an asterisk in an audit.
Covered in Module 05
Quarter review: 60 improvement actions opened, 14 closed, 9 verified in the ledger. A manager calls it 'a great quarter for the improvement pipeline.'
Q12 What does the open-vs-verified picture actually say?
A
Open even more actions — with more attempts, more will verify by volume alone. More opens onto an unclosed pile makes the ratio worse. What distinguishes a well-run plant is the rate at which actions close and verify.
B
A great quarter — a big pipeline means the plant is engaged and finding opportunity. Engagement that doesn't close and verify banks nothing. Sixty open items is sixty commitments the board now has to carry — or watch quietly rot.
C
Warning sign — the rhythm is generating activity, not banked results. Close rate and verified rate are the metrics to drive next quarter. Opening actions is easy. A growing open pile over a thin verified ledger means the cadence is producing meetings — the fix is follow-through, not more ideas.
Covered in Module 05